Where it is raised
- Errors ingest (/api/errors/v1/ingest)
- Errors feedback (/api/errors/v1/feedback), where the limit is 20 reports a minute
What triggers it
Three limits send this code: a key may send 300 batches a minute, a project has an events-per-minute ceiling that can take a whole batch, and a feedback key may send 20 reports a minute. The first is checked before the body is read, so a refused request costs no parsing. Unknown keys are counted per address.
Example response
POST /api/errors/v1/ingest/<key>
HTTP 429
Retry-After: <seconds>
{
"error": "Too many requests for this ingest key.",
"reason": "rate_limited",
"code": "RU-4003"
}Common causes
- An error storm where every request reports.
- More events in one minute than the project's limit; Pro and Business can raise it in Errors settings.
- SDK batching disabled, sending one event per request.
How to fix it
- Honor the Retry-After header.
- Enable SDK batching so many events ride one request.
- Fix the erroring code path: the rate limit is usually a symptom of an error storm.